An Optimal Algorithm for Stochastic and Adversarial Bandits.
Julian ZimmertYevgeny SeldinPublished in: AISTATS (2019)
Keyphrases
- dynamic programming
- optimization algorithm
- optimal solution
- exhaustive search
- theoretical analysis
- experimental evaluation
- worst case
- globally optimal
- improved algorithm
- monte carlo
- search space
- locally optimal
- learning algorithm
- times faster
- high accuracy
- significant improvement
- segmentation algorithm
- clustering method
- preprocessing
- matching algorithm
- closed form
- computational complexity
- optimal path
- stochastic approximation
- operating point
- weighting coefficients
- optimal control
- computationally efficient
- expectation maximization
- np hard
- k means
- search algorithm
- decision trees